Democracy: Should We Replace Elections with Random Selection?

In: Danish Yearbook of Philosophy
Annabelle Lever Professor of Political Philosophy, Sciences Po, Paris, France, 450819, Permanent Researcher, CEVIPOF, Paris, France

Search for other papers by Annabelle Lever in
Current site
Google Scholar
Open Access


The aim of this paper is to explore the claim that lotteries are more democratic than elections. The paper starts by looking at the two main forms of equality that give lotteries their democratic appeal: an individually equal chance to be selected for office, and the proportionate representation of groups in the legislature. It shows that they cannot be jointly realized and argues that their egalitarian appeal is more apparent than real. Finally, the paper considers the democratic reasons to value randomly selected assemblies, even if claims about their distinctively egalitarian properties are exaggerated.

Democracy and the Ethics of Voting

Over the past ten years or so, a lively debate has developed about the modalities of democratic elections and, to a lesser extent, the considerations that should guide citizens when voting. However, just as the ethics of voting has become an interesting subset of the political theory literature on democracy, another new literature suggests that voting in elections is not really a democratic way to select people for positions of power and authority at all. Instead, elections should be replaced in whole or in part by random selection or ‘sortition.’1 If this is correct, there would be no gain to democratic theory or practice in considering whether voting should be secret, mandatory, or strategic, and whether it should be focused on the common good of one’s country rather than on other moral or political considerations.2 The aim of this paper, therefore, is to explore the claim that lotteries are more democratic than elections. The paper starts by looking at the two main forms of equality that give lotteries their democratic appeal and shows that they cannot be jointly realized. Indeed, as we will see, even taken individually, their egalitarian appeal is more apparent than real. Finally, the paper considers the democratic reasons to value randomly selected assemblies, even if claims about their distinctively egalitarian properties are exaggerated.

Sortition and Democratic Equality

In The Principles of Representative Government, Bernard Manin reminded his readers that the association of representative government with democracy is relatively new, not simply because elections with universal suffrage and equally weighted votes are a recent phenomenon, but because the ancient and Renaissance republics considered elections to be an aristocratic, not a democratic, way to select people for political office.3 Elections were considered aristocratic because they enabled voters to select the people that they thought best qualified to rule, and to choose them again and again, if they so wished. By contrast, Manin explained, lotteries were thought of as democratic selection devices because they gave everyone the same chance of holding office and, when repeated, were likely to favor the rotation of people in power, thereby militating against the creation of a political elite. However, until recently no one seems to have concluded that elections, even with universal suffrage, are less democratic than sortition, or that it would be a gain for democracy if people replaced one or more of their legislative bodies with a chamber created by sortition, or random selection.4

So, what can be said in favor of the democratic credentials of sortition—bearing in mind that lotteries, often in conjunction with elections, were used to prop up oligarchical regimes in the Renaissance republics of Venice and Florence and, even as late as the nineteenth century, in Swiss cantons like Berne?5 The obvious point, and one repeatedly cited by contemporary fans of random selection, is the idea that lotteries, when suitably constructed, answer to the democratic idea that people are equally entitled to participate in politics, and to take part in ruling, as well as being ruled. The equal opportunity to hold office created by unweighted lotteries, as well as the rotation in office that they promise, seems to speak directly and in an intuitively appealing way to the idea that citizens are each entitled to take part in the government of their society and are, in principle, equal, even interchangeable, in their claims on office.6 When it is impossible to share a good equally, random selection among claimants preserves the commitment to treating them as equals by avoiding invidious, destructive, and unfair comparisons among them.7 Importantly, unequal rewards justified by lot rather than choice do not impugn the virtues, capacities, status, needs, or desires of those who lose out, and therefore provide no grounds for arrogance or preening on the part of winners, or self-abasement on the part of losers.8 In that sense, the inequality of rewards created by lotteries seem more likely to maintain people’s sense of themselves as peers than the inequality of rewards based on choice. Hence, it might seem that unweighted lotteries are the democratically ideal way to distribute goods, including political office, to which all have equal claim but which it is impossible or undesirable to share among everyone at once.

Some additional considerations about randomization are also relevant to recent arguments in favor of its democratic credentials. Where the randomly selected body is sufficiently large relative to the total population, and where everyone selected participates, unweighted lotteries create assemblies that are a microcosm of the larger population. That is, the randomly selected assembly, though smaller than the whole population, will be made up of different groups in proportion to their numbers in the population as a whole.9 This will be true of the invisible as well as the visible attributes of citizens, and it will therefore be possible to treat the smaller group as an exact replica of the larger one, and as an accurate replacement for it for certain purposes. Although randomization with smaller bodies is likely to result in clusters, meaning that the smaller group will not be a microcosm of the larger one, as long as the assembly is large enough relative to the total population such problems will disappear—for the same reason that, over a large enough number of coin tosses, heads will fall face up 50 percent of the time, although for smaller stretches of time tails may dominate, or vice versa. Randomization then can be compatible with microcosmic selection, or ‘mirror representation,’ as Pitkin called it, and ‘descriptive representation,’ as it is now usually called.10

Mirror representation once seemed at odds with the uniformity, not just the universality, required for democratic rights, even by those who believed it essential that workers should be enfranchised.11 For example, proponents of mirror representation worried that enfranchising workers would create a legislative body in which smaller groups in the population could be ignored in legislative deliberations, because they would be consistently outvoted in Parliament. However, mirror representation also appeared at odds with democratic representation because proponents of the former tended to treat the individual members of social groups as interchangeable, and therefore saw no reason why, qua members, they all needed to have the same political rights as each other.

Replacing elections with lotteries, however, suggests that we can combine democracy and mirror representation, reconciling uniform and equal rights to be selected, on the one hand, with outcomes that replicate the diversity of the population in the legislature, on the other hand. Thus, while randomization means that many—perhaps most of us—will not be part of a legislative assembly, we will know that we had the same chance to be selected as everyone else, and that despite our absence, we will be represented by others who think, feel, and live like us in numbers that fairly represent our frequency in the population.12 In contrast, most elected legislatures suffer from forms of descriptive misrepresentation that, since the 1990s, have come to be seen as evidence of the continued weight of structural inequalities on people’s political opportunities and ambitions, and therefore as undermining the ability of governments to claim that they speak for all of us.13 Random selection, then, seems to have a democratic appeal that elections lack, enabling us to combine an individually equal opportunity for office with the proportionate representation of individuals as members of distinctive, even competing, groups.

Unweighted Lotteries versus Substantive Representation

Unfortunately, for reasons both of relative size and of people’s willingness and ability to participate, it is impossible to create deliberative assemblies that preserve both an equal opportunity for individuals to be selected and a microcosmic or descriptively representative character. Unweighted lotteries standardly create assemblies that do not look at all like the citizen body—both because they are too small to ensure that different groups of citizens are represented proportionately, or even at all, and because citizens are often unwilling and/or unable to participate when selected.14

Both problems are non-trivial. Evidence from the larger citizen assemblies, such as the Scottish assembly on climate change, suggests that only 3 percent of those selected randomly respond positively to the invitation to take part, and these assemblies are normally made up of only 150 to 250 people. Much higher proportions of favorable responses have been found among those randomly selected for smaller and less onerous assemblies of about 50 people, but the average positive response rate for all assemblies is currently a bare 15 percent of those invited.15 We may hope, with Alexander Guerrero, that high salaries, relocation assistance, and other forms of support and encouragement might persuade people selected at random to take on the onerous responsibilities of a legislative body.16 However, realism suggests that most people are unlikely to want to devote large amounts of their time to politics, even for a relatively finite period, or to disrupt their professional and private relationships to shoulder public responsibilities for which they may have no taste or affinity. In such circumstances, formally equal opportunities for individuals come at the cost of the equal representation of groups, and in particular of people’s membership in ascriptive groups, conceived as an important cause of their differential life-chances, status, and resources.

So, it seems that we must choose between the two forms of equality that make random selection look democratically appealingly, as we cannot have both. Under modern circumstances—where populations are too large for everyone to participate together in a lawmaking assembly, as in Athens—it is impossible to see how equal opportunity among individuals and the equal representation of groups are to be preserved, for the same reasons that short runs of coin tosses with tails on top are statistically possible, even probable. The problem is exacerbated, but not created, by current practices of voluntary, rather than mandatory, participation in politics, because the unequal representation of groups that voluntariness creates is likely to be skewed heavily toward those who are self-confident and politically motivated, and those with the time and resources to devote themselves to political, rather than to personal or professional, matters. Thus, the impartiality that gives randomization its egalitarian appeal looks significantly less democratic once formally equal opportunities to serve become the means through which inegalitarian life-chances and resources are transmuted into descriptively unrepresentative legislatures.

Unweighted Lotteries and the Challenges of Equality

By design, a randomly selected legislature is one that has been selected without input from those it governs. A legislative body created by such means has a democratic appeal that a randomly selected monarch lacks only because it is an assembly and therefore capable of representing the different beliefs, needs, and interests of the population in ways that a single person cannot. The lack of citizen agency implicit in the random selection of governors, then, is supposed to be compensated by the fact that they look like us and can be expected to have our best interests at heart collectively, if not individually. As Peter Stone puts it: “Elected representatives can say, we represent you because you chose us. Randomly-selected representatives can say, we represent you because we are indistinguishable from you. (The latter, but not the former, is inextricably a collective claim. It is hard for a single randomly-selected citizen to claim to “represent” anybody.)”17

Unweighted lotteries, however, undermine that appeal, because they provide no grounds to believe that those people who were selected randomly and are also willing and able to serve will be a sufficiently diverse bunch to understand and sympathetically articulate and promote our different needs and interests. Worse, it may turn out that their interests and beliefs conflict systematically with our own, in ways that would have been less likely had we selected a monarch rather than an assembly. Just as troubling if we care about democratic equality, however, is the fact that we will be in no position to re-select an assembly that, through good luck, happens to work well for us all and to reflect our views. Given the descriptive misrepresentation that unweighted lotteries are likely to create, therefore, it is difficult to say with confidence that an assembly created by such means is better than a randomly selected monarch, let alone that it constitutes an egalitarian improvement over elections, whatever the problems with the latter.

The rationale for using unweighted lotteries to create a legislative assembly is, as we have seen, distributive: it gives everyone an equal chance of being selected for office. Hence, it is indifferent to citizen agency as a component of democratic ideas of equality—not just in the choice of representatives, but with regard also to agenda-setting and to legislative accountability.18 However, even from a purely distributive perspective, the egalitarian case for favoring lotteries over other ways of selecting a government is unpersuasive.

In ancient Athens, where all citizens were entitled to take part in the primary political body (which, prior to the reforms of 403/2 bce that initiated the distinction between laws and decrees, was responsible for passing both), unweighted lotteries were used to select from a pool of volunteers for supplementary offices.19 Lotteries, therefore, were used to distribute a scarce good that volunteers desired, and to which they were thought to have equal claims. In those circumstances, lotteries may have been a fair way to treat equals, just as they can be a fair way to distribute burdens among a group of people who are equally liable to bear them.20 However, when we are dealing with a population who disagree about the value of the good to be distributed, in that some desire it whereas others see it as a burden, the egalitarian rationale for random selection vanishes. Giving everyone the same opportunity to be selected for office, regardless of their attitude to office-holding, is likely to seem unmotivated and confusing, at best. At worst, as Michael Walzer worried, such indifference to the rationale for equality—to who it is supposed to benefit, or what it is supposed to achieve—renders the ideal of equality “ripe for betrayal.”21 Strikingly, Hélène Landemore and David Owen and Graham Smith are willing to consider compulsory service in their randomly selected legislative bodies,22 while Arash Abizadeh and Guerrero reject that possibility.23 Yet, their claims on behalf of random selection are very similar, although Guerrero is primarily focused on the epistemic advantages of lottocracy, and might therefore have been expected to look more favorably on compulsion in principle, as compared to those for whom the justification of lotteries is essentially egalitarian and where, therefore, the significantly unequal burdens of participation can be expected to militate against compulsion.

If we see representative office as a good to be distributed fairly among those who are equally entitled to it, there is no reason to make participation in the lottery itself involuntary, or to suppose that those selected should be forced to serve if an unexpected turn of events makes it difficult. We may think it important to publicize widely the attractions of office and the deadlines and other requirements to take part in the draw, and may want to make special efforts to recruit and support people from disadvantaged social groups so that they, too, are able to share in “the realization of self which comes from a skilful and devoted exercise of social duties”—what Rawls referred to as “one of the main forms of the human good.”24 If we are all equally capable of ruling over others, then, there will be no paternalistic rationale for forcing people to take part in the lottery or to serve if they do not want to;25 nor will there be an egalitarian justification for doing so as long as the lottery is concerned with the fair distribution of a scarce resource.

Conversely, if we think that office-holding is a burden—that its moral and physical demands are likely to be frightening, difficult, perhaps overwhelming—we will expect that participation in the lottery be mandatory and that people be required to serve if selected, with few exceptions. Compulsion, here, is the egalitarian corollary of inclusion as equals, each functioning together to ensure that unwanted burdens are fairly shared by all.26 In short, while a randomly selected legislature might answer to concerns for the fair sharing of benefits or burdens among equals, the same cannot be said when random selection gives us an identical chance at office whether we want it or not.

If those included in the draw for office involuntarily are entitled to refuse to serve with no need for self-justification, unwanted publicity, or bureaucratic entanglements, they may suffer no harm from their unequal treatment—as compared to most regimes of compulsory voting.27 But they may feel that they have been wronged, because the burdens of fairly deciding among volunteers for office falls almost solely on those who do not want to volunteer. Public-spirited, capable citizens who do not want or are unable to hold office, then, may object to a form of lottery that includes them in an effort to select among volunteers, thereby implying that they do not exist, or do not matter. More seriously, they may wonder whether the choice of random selection over other ways of distributing offices adequately reflects their status and claims as citizens. After all, they are generally willing to select among candidates for office, even if unwilling and/or unable to take up office themselves.28 Yet neither their willingness nor its limits plays a role in determining what equality means in access to legislative office.

In short, the egalitarian appeal of unweighted lotteries for office does not withstand critical scrutiny. There is nothing egalitarian about giving the willing and unwilling the same shot at office. Nor will there be much to say for democracy if political office is so unappealing that it must be imposed, however fairly, on the unwilling. Finally, randomizing among self-selected volunteers is not a good way to constitute a democratic legislature. Hence, we cannot generalize the egalitarian appeal of unweighted lotteries to the constitution of democratic legislatures.

Equality, Weighted Lotteries, and the Claims of Individuals as Members of Groups

Unweighted lotteries lack the egalitarian benefits that their proponents claim for them. It is not surprising, then, that most citizen assemblies are constructed using weighted lotteries, to try to equalize representation in the legislature of disadvantaged and advantaged groups, and to ensure that the vast majority of the population who will not be selected for office can, nonetheless, have a legislature that looks like them and/or is likely to think as they would.

Of course, using weighted lotteries to construct a legislative assembly means that we will have to give up the claim that our assembly is democratic because it gives everyone the same chance to be selected for office. As we now know, this concession is more likely to enhance rather than detract from the egalitarian case for lotteries. However, clarity about the trade-off between formal equality of opportunity and the use of weighted lotteries is uncommon among proponents of lotteries, where it is still customary to assert that sortition assemblies are both descriptively representative and examples of formally equal individual opportunities for office.29

At present, the stratified sampling used to create citizen assemblies is generally based on characteristics such as age, sex, education, geography, and income. These characteristics, it is plausibly assumed, are politically pertinent, in that these factors tend to group citizens into ascriptive and voluntary groups in which members can be expected to have shared interests in common, as well as ones that may bring them into conflict with other citizens. In the United Kingdom it is common also to stratify based on ethnicity, given that this is a pertinent and often politically divisive social characteristic, creating hierarchies of power, privilege, and opportunity that are largely unchosen.30 However, in Europe, ‘ethnic statistics’ are often illegal, as the effort to group people based on their real or supposed ties of language, birth, and culture is considered likely to promote, rather than to combat, unfair distinctions among citizens.31 Consequently, in Europe citizen assemblies are unlikely to reflect ethnic or racialized groups in proportion to their numbers in the population, important though such factors may be in explaining and justifying people’s political beliefs and opinions. In addition to differences in this respect, the climate assemblies in the United Kingdom, as distinct from the one in France, were also stratified based on people’s opinion on climate change. Thus, the British climate assemblies sought to include those in denial about climate change, as well as those convinced by, and horrified at, its gravity.32

Weighted lotteries, then, look like an egalitarian response to the challenges of voluntary participation under circumstances of structural inequality that link past disadvantage to unequal rates of political participation in the present. Understood thus, the case for weighting is, essentially, egalitarian. However, weighting is likely to be justified even where structural inequality is not an issue, given the very small size of any deliberative assembly compared to the population as a whole. In this case, the goal is to ensure that individuals are adequately represented in the legislature, given that politics is both a competitive and cooperative business and that adequate representation of the whole population will not happen automatically.

Hence, in both ideal and non-ideal circumstances, the use of weighted lotteries implies hard, and perhaps politically controversial, choices about which groups are to be represented in a legislative assembly, and in what proportion relative to each other. These choices are likely to be particularly difficult for an egalitarian because, as Melissa Williams has put it, “no defensible claim for group representation can rest on assertions of the essential identity of women or minorities; such assertions do violence to the empirical facts of diversity as well as to the agency of individuals to define the meaning of their social and biological traits.”33 A more serious problem, however, is that weighted lotteries cannot prevent the very significant gap between those selected and those serving from undermining the egalitarian appeal of sortition assemblies, however much they may look like us.

For example, the organizers of the Climate Assembly UK sent out 20 percent of their 30,000 letters of invitation to people randomly selected from the lowest-income postcodes, and then used random stratified sampling by computer to select 110 participants from all the people who were over 16 and free on the relevant dates. Or consider Philippe Van Parijs’s surprise that it took 50,000 phone calls to get 700 or so people to turn up for a deliberative assembly, the G1000 in Brussels, on November 11, 2011.34 Such an extreme disparity between those called and those serving precludes confidence in the moral, political, or epistemic qualities of the latter. Specifically, the fact that those selected and serving are, to that extent, more like each other than like the rest of the population, makes it hard to be sure that they have the internal diversity necessary to embody the wisdom, impartiality, and range of experiences required for their egalitarian appeal as a body.35 In short, it remains to be seen what egalitarian virtues or properties we can attribute to assemblies created by weighted sortition.

Two further problems are worth noting. Most randomly selected assemblies, as we have seen, make a special effort to ensure the presence of women in their bodies, roughly in proportion to their number in the general population. But although the organizers will have made an effort to represent the interests and beliefs of women qua women, on the assumption that these may well be different from, and in conflict with, the interests of men, no effort is ever made to ensure that the politically relevant differences among women will be adequately represented too. However, the idea that women can be represented politically while abstracting from differences of faith, experience, and affiliation that make them the type of women they are has long been rejected as fostering an inappropriate essentialism about women and an indifference to the significance of intersectionality for their experiences as women.36 Put otherwise, we cannot ensure that women are politically represented as equals if they must count on men to represent the differences of age, income, occupation, education, wealth, religion, race, ethnicity, and gender that may set them in conflict with other women.

Above all, if we care about democratic equality, it is unclear why we should take the size and group distribution of the existing population as a model of what an ideal assembly should look like once we are in the business of consciously creating our legislatures. The composition of national populations, after all, is the result of many factors that are morally unappealing. These include harsh, humiliating, and discriminatory immigration regimes and denials of citizenship37 that, however popular, reinforce ideals of citizenship that are sexually, racially, and ethnically inegalitarian. Our population, as well, is affected by rates of murder and other forms of violence, as well as of unequal access to healthcare, housing, nutrition, education, and jobs that affect how long we live as well as the fate of our children. As Anne Phillips has emphasized, “The underlying preoccupation [with group-sensitive representation] is not with pictorial adequacy—does the legislature match up to the people?—but with those particularly urgent instances of political exclusion which a ‘fairer’ system of representation seeks to resolve.”38 Thus, an egalitarian perspective on random selection is as likely to challenge, as to affirm, the unstated assumption dominating contemporary theory and practice—namely, that we should seek to represent politically salient groups in our legislatures in proportion to their numbers in the population.

As we know, disproportionate representation may be necessary for small and disadvantaged groups to be heard in the legislature, rather than ignored or consistently outvoted on matters that concern them deeply. Citizen assemblies made up predominantly, perhaps exclusively, of disadvantaged groups may be necessary to combat forms of “cultural imperialism,” as Iris Young calls it39—that combination of being silenced, devalued, and judged by the standards, experience, and knowledge of the more privileged. While well-meaning, then, the effort to secure two places for those in ‘deep poverty’ in the French climate assembly was sheer tokenism,40 unless efforts were made to ensure that they could effectively articulate what climate change, and different strategies to combat it, might mean to them. Driving on the highway at one speed limit rather than another, eating meat rather than other things, or taking the plane rather than the train may be matters on which the extremely disadvantaged have views—and ones that they are willing to share with others—including the urgency of ‘significantly modifying one’s lifestyle’ in order to prevent climate change.41 But if it matters that an assembly understands what climate politics look like to the most disadvantaged, we will need a more complex conception of equality than that used by the organizers of the French Convention for the Climate, or than is common in the literature on sortition assemblies.42


Sortition has many virtues as a way to distribute benefits and burdens among equals where rotation or sharing are neither possible nor desirable. Its unpredictability makes favoritism difficult, even impossible—and this can be true even when weighted lotteries are used to distribute benefits and burdens among a large group of people. Unweighted lotteries give everyone the same chance of being selected, and that can be appealing if it is publicly important to insist on an inclusive conception of equality before the law in the face of past prejudice, second-class status, and exclusion. Weighted lotteries enable the patterning of outcomes, which can be appealing when formal equality of opportunity is less significant democratically than the ability to combat those forms of marginalization, exclusion, coercion, and cultural imperialism that formally equal political rights do not touch. In short, lotteries can be a helpful way to distribute scarce goods and opportunities, as well as unwanted burdens, when self-selection alone, or in conjunction with selection by others, is likely to be unfair.43

Moreover, randomly selected assemblies are democratically attractive in so far as they enable citizens to deliberate with strangers as equals on matters of collective importance, with the help and support of a variety of experts. Given how scarce such opportunities are, and how seldom we get to discuss politically relevant information in a critical but supportive environment, such randomly selected assemblies are invaluable, especially for those who are not members of an organized public body. Where consensus is impossible, assemblies can clarify the nature and depths of disagreement, which is democratically valuable in itself and may help to structure future legislative debates and decisions constructively. At present, the norm is to produce a single assembly report made up of recommendations that all could accept, but there is no reason why this should be the case. Where deliberative consensus among randomly selected citizens is possible, and substantiated over time, it can enable politicians to take difficult decisions that they might otherwise be reluctant to make and push them to recognize the gap between their claims to reflect and serve the public and the reality of their distance from it. In short, citizen assemblies can be occasions for democratic experimentation, expression, solidarity in disagreement, inspiration, and hope for better forms of politics, even as advisory bodies.

However, when it comes to the two forms of equality that were meant to give sortition assemblies their distinctively democratic appeal—the equal chance to be selected, and the equal representation of groups—their advantages have proved elusive. The defects in the egalitarian case for lotteries, which we have examined, are real. But their interest, I believe, lies less in what they tell us about the relative merits of elections and lotteries as democratic selection devices, and more in what they reveal about the challenges of democratic equality, no matter how we construct our legislatures. Recognizing and meeting those challenges requires moving beyond a preoccupation with the deficiencies of electoral democracy as manifested in the United States, common in the work of Landemore, Abizadeh, and Guerrero, to an engagement with the different forms that electoral democracy can take, and the different ways in which it can be combined with forms of direct democracy at the local level, with referenda from above and below, and even with “aleatory democracy,” as Stone calls it.44 Finally, the difficulties of the egalitarian case for lotteries remind us that current practices for constituting and running sortition assemblies are in their infancy. It would be a significant moral, political, and epistemic mistake to take its current theory and practice as a ‘gold standard’ for deliberative democracy,45 thereby discouraging much-needed reflection and experimentation in future.


Many people have read or heard previous versions of this paper. I owe special thanks to Chiara Destri, Attila Mraz, Peter Stone, Julian Culp, Alexu Volacu, and the editors of this journal for their critical comments and constructive suggestions. Above all, I thank Antje Gimmler, Jorn Sonderholm, and the Faculty of Social Sciences and Humanities at Aalborg University for inviting me to give the lecture on which this paper is based. Research for this paper was funded by the Horizon 2020 Framework Programme, Project 870996, “Reconstructing Democracy in Times of Crisis: A Voter-Centred Perspective.”


Arash Abizadeh, “Representation, Bicameralism, Political Equality, and Sortition: Reconstituting the Second Chamber as a Randomly Selected Assembly,” Perspectives on Politics 19, no. 3 (2021): 791–806; Hélène Landemore, Open Democracy: Reinventing Popular Rule for the Twenty-First Century (Princeton, NJ: Princeton University Press, 2020); oecd, Innovative Citizen Participation and the New Democratic Institutions: Catching the Deliberative Wave (Paris: oecd Publishing, 2020); David Owen and Graham Smith, “Sortition, Rotation, and Mandate: Conditions for Political Equality and Deliberative Reasoning,” Politics and Society 46, no. 3 (2018): 419–434; Yves Sintomer, “Sortition and Politics: From Radical to Deliberative Democracy—and Back?,” in Brill’s Companion to the Reception of Athenian Democracy, ed. Dino Piovan and Giovanni Giorgini (Leiden: Brill, 2020), 490–521; Graham Smith, Can Democracy Safeguard the Future? (Cambridge: Polity Press, 2021).


Geoffrey Brennan and Philip Pettit, “Unveiling the Vote,” British Journal of Political Science 20, no. 32 (1990): 311–333; Jason Brennan, The Ethics of Voting (Princeton, NJ: Princeton University Press, 2012); Brennan, “Polluting The Polls: When Citizens Should Not Vote,” Australasian Journal of Philosophy 87, no. 4 (2009): 535–549; Bart Engelen, “Against the Secret Ballot: Toward a New Proposal for Open Voting,” Acta Politica 48, no. 4 (2013): 490–507; Engelen, “Why Liberals Can Favour Compulsory Attendance,” Politics 29 (2009): 218–222; Annabelle Lever and Alexandru Volacu, “Should Voting Be Compulsory? Democracy and the Ethics of Voting,” in The Routledge Handbook of Ethics and Public Policy, ed. Annabelle Lever and Andrei Poama (London: Routledge, 2018), 242–254; Lever, “Must We Vote for the Common Good?,” in Ethics in Politics, ed. Emily M. Crookston, David Killoren, and Jonathan Trerise (New York: Routledge, 2016), 145–156; Lever, “Mill and the Secret Ballot: Beyond Coercion and Corruption,” Utilitas (2007): 354–378; Lachlan Umbers, “Compulsory Voting: A Defence,” British Journal of Political Science 50, no. 4 (2020): 1307–1324.


Bernard Manin, The Principles of Representative Government (Cambridge: Cambridge University Press, 1997).


Abizadeh, “Representation”; Hubert Buchstein, “Reviving Randomness for Political Rationality: Elements of a Theory of Aleatory Democracy,” Constellations: An International Journal of Critical and Democratic Theory 17 (2010): 435–454; Claudia Chwalisz, “Eight Ways to Institutionalise Deliberative Democracy,” oecd Public Governance Policy Paper, 2021; Alexander A. Guerrero, “The Epistemic Pathologies of Elections and the Epistemic Promise of Lottocracy,” in Political Epistemology, ed. Elizabeth Edenberg and Michael Hannon (Oxford: Oxford University Press, 2021), 156–179; Guerrero, “Against Elections: The Lottocratic Alternative,” Philosophy and Public Affairs 42, no. 2 (2014): 135–178; Landemore, Open Democracy; David Van Reybrouck et al., The Malaise of Electoral Democracy and What to Do About It, Re-Bel e-book 14 (April 2014),


Maxime Mellina et al., Tirage au sort et politique: Une histoire suisse (Lausanne: PU Polytechnique, 2021); Gil Delannoi, Le tirage au sort: Comment l’utiliser? (Paris: Presses de Sciences Po, 2019).


Owen and Smith, “Sortition”; Landemore, Open Democracy, 90.


Ben Saunders, “The Equality of Lotteries,” Philosophy 83 (2008): 359–372.


Delannoi, Le tirage au sort, 95–97; Montesquieu, L’esprit des lois (1748), bk. 2, ch. 3.


Gil Delannoi and Oliver Dowlen, Sortition: Theory and Practice, vol. 3 (Luton: Andrews UK Limited, 2016), 14; Landemore, Open Democracy, 90.


Hannah Fenichel Pitkin, The Concept of Representation (Berkeley: University of California Press, 1967).


Gregory Conti, Parliament the Mirror of the Nation: Representation, Deliberation, and Democracy in Victorian Britain (Cambridge: Cambridge University Press, 2019), 27–76.


Guerrero, “Against Elections,” 167.


Jane Mansbridge, “Should Blacks Represent Blacks and Women Represent Women? A Contingent ‘Yes,’” Journal of Politics 61, no. 3 (1999): 628–657; Anne Phillips, The Politics of Presence (Oxford: Oxford University Press, 1995); Melissa S. Williams, Voice, Trust, and Memory: Marginalized Groups and the Failure of Liberal Representation (Princeton, NJ: Princeton University Press, 1998).


Vincent Jacquet, Comprendre la non-participation: Les citoyens face aux dispositifs délibératifs tirés au sort (Brussels: Peter Lang, 2020); Jacquet, “Explaining Non-Participation in Deliberative Mini-Publics,” European Journal of Political Research 56, no. 3 (2017): 640–659.


Jean-Michel Fourniau, “La sélection des mini-publics entre tirage au sort, motivation et disponibilité,” Participations, special isssue, “Tirage au sort et démocratie: Histoire, instruments, théories” (2019): 373–400; Jean-Michel Fourniau, Bénédicte Apouey, and Solène Tournus, “Le recruitement et les caractéristiques sociodémographiques des 150 citoyens de la Convention citoyenne pour la climat,” working paper (Paris, 2020),


Guerrero, “Against Elections,” 156; Guerrero, “Epistemic Pathologies,” 170.


Peter Stone, “The Two Faces of Sortition,” unpublished paper, 10.


Emmanuela Ceva and Valeria Ottonelli, “Second-Personal Authority and the Practice of Democracy,” Constellations: An International Journal of Critical and Democratic Theory 29, no. 4 (2021): 460–474; Chiara Destri and Annabelle Lever, “Egalité démocratique et tirage au sort,” Raison Publique, forthcoming; Cristina Lafont, Democracy without Shortcuts: A Participatory Conception of Deliberative Democracy (Oxford: Oxford University Press, 2020); Lachlan Umbers, “Against Lottocracy,” European Journal of Political Theory 20, no. 2 (2021): 312–334. Peter Stone refers to this indifference to agency in “Popular Rule without Popular Sovereignty,” his critique of Landemore’s Open Democracy, forthcoming in Democratic Theory (2024). His article makes clear why attitudes to popular sovereignty mark a critical difference between those who see lotteries as the best expression of democratic ideals and those for whom referenda, elections, or other expressions of popular will and agency are also required.


Many thanks to Alexandru Volacu for highlighting the significance of the reforms of 403/2 for these issues.


Saunders, “Equality of Lotteries.”


Michael Walzer, Spheres of Justice: A Defense of Pluralism and Equality (New York: Basic Books, 1983), xi.


Owen and Smith, “Sortition,” 431, where the absence of compulsion would justify demographic stratification.


Abizadeh, “Representation,” 10; Guerrero, “Against Elections,” 156; Guerrero, “Epistemic Pathologies,” 170.


John Rawls, A Theory of Justice (Cambridge, MA: Belknap Press of Harvard University Press, 1971), 1.14, p. 84.


Compare Hill, Birch, Lijphart, and Engelen, who favor compulsory voting in large part because they believe it will benefit the socially disadvantaged, who are disproportionately unlikely to vote. Lisa Hill, “Low Voter Turnout in the United States: Is Compulsory Voting a Viable Solution?,” Journal of Theoretical Politics 18, no. 2 (2006): 207–232; Hill, “On the Justifiability of Compulsory Voting: Reply to Lever,” British Journal of Political Science 40 (2010): 917–923; Sarah Birch, Full Participation: A Comparative Study of Compulsory Voting (Manchester: Manchester University Press, 2008); Arend Lijphart, “Unequal Participation: Democracy’s Unresolved Dilemma,” American Political Science Review 91 (1997): 1–14; Engelen, “Why Liberals”; Annabelle Lever, “Compulsory Voting: A Critical Perspective,” British Journal of Political Science 40, no. 4 (2010): 897–915; Lever, “Democracy and Voting: A Reply to Lisa Hill,” British Journal of Political Science 40, no. 4 (2010): 925–929; Lever and Volacu, “Should Voting Be Compulsory?”


Jury service typically randomizes across people who do and do not want to serve. Entry into the draw is involuntary and one is presumptively obliged to serve if chosen. However, the point of compulsion in this case is fairness to defendants and victims of crime, rather than fairness to other citizens. I therefore differ from the account of jury lotteries given by Peter Stone (“Sortition, Voting, and Democratic Equality,” Critical Review of International Social and Political Philosophy 19, no. 3 [2016]: 350), which focuses on the right to be a juror, but which ignores the reasons why jury service is a duty. Fairness to actual and potential jurors explains why exemptions of the competent are sometimes justified, and why court boundaries should be drawn so that jurors do not need to travel excessively. But as the principal obligation of jurors is fairness to defendants and victims of crime, the selection of jurors has that as its primary focus, as does court procedure more generally.


Cf. Birch, Full Participation.


Being willing to act as a voter, in principle, is different from being prepared to vote in all elections. Indeed, the reasons to value the right to vote may be reasons to insist on the right not to vote in certain circumstances—for example, when there was no reason to call an election now, or to protest the quality of the candidates/programmes on offer. Lever, “Compulsory Voting.”


See, e.g., Abizadeh, “Representation,” 798. For example, Alessandro Bellatoni, Claudia Chwalisz, and Leva Cesnultaityte, for the oecd, claim that “The participants should be a microcosm of the general public. This is achieved through random sampling from which a representative selection is made,” and then require that “Everyone should have an equal opportunity to be selected,” as Point 5 in their guide to good practice. See Bellatoni et al., “Good Practice Principles for Deliberative Processes for Public Decision Making,” 6 ( This confusion pervades oecd publications on random selection and ‘the deliberative wave,’ reappearing in their accounts of ‘good practice principles.’ oecd, Innovative Citizen Participation, 118. See also Claudia Chwalisz, “A Movement That’s Quietly Reshaping Democracy for the Better,” Noēma, May 12, 2022, “The civic lottery needs to ensure that everyone has an equal chance of being selected,” yet a group of one hundred Parisians is “broadly representative of the community,” “a microcosm of the wider public.”


The Council of Europe, however, recognizes that racial and ethnic discrimination can make relevant statistics a necessary tool in understanding and fighting discrimination. See “‘Ethnic’ Statistics and Data Protection in the Council of Europe Countries: A Study Report,” October 31, 2007, However, as Patrick Simon, a French specialist on the topic has noted, the use of such statistics remains uncommon in Western Europe. Patrick Simon, “Collecting Ethnic Statistics in Europe: A Review,” Ethnic and Racial Studies 35, no. 8 (2012): 1366–1391; Simon, “The Failure of the Importation of Ethno-Racial Statistics in Europe: Debates and Controversies,” Ethnic and Racial Studies 40, no. 13 (2017): 2326–2332.


Convention Citoyenne pour le Climat,; Climate Assembly UK,; “Scotland’s Climate Assembly—Process, Impact and Assembly Member Experience: Research Report,”


Williams, Voice, Trust and Memory, 6. Anne Phillips quotes Williams approvingly in “Descriptive Representation Revisited,” in The Oxford Handbook of Political Representation in Liberal Democracies, ed. Robert Rohrschneider and Jacques Thomassen (Oxford: Oxford University Press, 2019), 174–191.


Van Parijs also notes his surprise that “what emerged to the outside world as the tangible outcome of the event reflected only to a minute extent the discussion on which the day was spent.” As he puts it, “Neither of these facts destroys the relevance of the experiment. But each invites to modesty as regards its potential.” Van Reybrouck et al., Malaise of Electoral Democracy, 51.


See Guerrero, Political Epistemology; compare Lafont, Democracy without Shortcuts, and Annabelle Lever, “Random Selection, Democracy and Citizen Expertise,” Res Publica (advance online publication, March 2023),


Kimberlé Crenshaw, “Demarginalizing the Intersection of Race and Sex: A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory, and Antiracist Politics,” in Living with Contradictions: Controversies in Feminist Social Ethics, ed. Alison M. Jaggar (Boulder, CO: Westview Press, 1994), 39–52.


For a recent, particularly controversial case from the UK, see Haroon Siddique, “Shamima Begum Loses Appeal against Removal of British Citizenship,” Guardian, February 22, 2023,


Phillips, “Descriptive Representation Revisited,” 47.


Iris Marion Young, Justice and the Politics of Difference (Princeton, NJ: Princeton University Press, 2011), 59–63.


Adrien Fabre et al., “Who Are the Citizens of the French Convention for the Climate?,” Paris School of Economic Working Paper No. 2021–39, April 2021, Of the 150 citizens who constituted the French ccc, “Five people from overseas collectivities were recruited among those studying in metropolitan France to avoid overseas journeys. In addition, two people in deep poverty were directly recruited through the association ‘Les petits frères des pauvres.’” A significant difficulty with this otherwise interesting comparison of the members of the ccc and their conclusions with that of the general French population is that the same statistical factors and weightings used to choose the former were used to sample the latter—thereby providing no guide to the adequacy of these categories as a guide to what French people think.


Fabre et al., “Who Are the Citizens?,” 11–14.


For efforts to think through what a fairer process might involve—but with a recognition of the challenges—see Graham Smith, “Reflections on xr Citizens’ Assembly Working Group Meeting on Decolonization,” July 26, 2019 (unpublished paper, on file with this author).


Saunders, “Equality of Lotteries”; Neil Duxbury, Random Justice: On Lotteries and Legal Decision-Making (Oxford: Oxford University Press, 1999); Peter Stone, The Luck of the Draw: The Role of Lotteries in Decision Making (Oxford: Oxford University Press, 2011); Barbara Goodwin, Justice by Lottery (Luton: Andrews UK Limited, 2013).


Alice el-Wakil, “The Deliberative Potential of Facultative Referendums: Procedure and Substance in Direct Democracy,” Democratic Theory 4, no. 1 (2017),; Alice el-Wakil and François Cheneval, “Designing Popular Vote Processes to Enhance Democratic Systems,” Swiss Political Science Review 24, no. 3 (2018): 348–358; Attila Mráz and Annabelle Lever, “Synthesis of Democratic Political Systems and Institutions in Europe” (2021) and “Voter-Centred Perspectives vs. Alternative Approaches to Electoral Democracy” (2021), both available online at; Maud Dugrand, La petite République de Saillans: Une expérience de démocratie participative (France: Editions du Rouergue, 2020); Peter Stone, “Democracy in Ireland: Theory and Practice,” in The Oxford Handbook of Irish Politics, ed. David M. Farrell and Niamh Hardiman (Oxford: Oxford University Press, 2021), 89–106.


The quest for a ‘gold standard’ figures prominently in recent oecd publications on ‘the deliberative wave.’ See, for instance, oecd, Innovative Citizen Participation, 32. The search for ‘good practice principles’ to be shared is appealing. The difficulty is the uncritical assumption that current practice adequately embodies such principles.

Content Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 604 604 119
PDF Views & Downloads 823 823 182